Summary
Erfan Shayegani is a PhD candidate in Computer Science at UC Riverside specializing in AI safety, multimodal understanding, and AR/VR security and privacy, with four years of research and industry experience. He has led high-impact work on cross-modality safety alignment for multimodal LLMs—first-author publications in ICLR, ICML, ACL, and EMNLP and a Best Paper Award at SoCal NLP—while interning twice at Microsoft Research where his projects on empathetic agents and Computer-Use Agents influenced product teams and executive attention. His research blends measurement, evaluation, and steering techniques (including unlearning and reward-model based approaches) and produced practical artifacts adopted by Azure AI services and a BLIND-ACT benchmark exposing “blind goal-directedness.” Comfortable moving between applied engineering and rigorous evaluation, he has prior experience building NLP-driven smart healthcare systems and teaching AI at Sharif University of Technology. Colleagues describe him as interdisciplinary, self-motivated, and adept at turning theoretical insights into deployable tools that inform both safety research and product development.
4 years of coding experience
2 years of employment as a software developer
Doctor of Philosophy - PhD, Computer Science, Doctor of Philosophy - PhD, Computer Science at University of California, Riverside
Bachelor's degree, Electrical engineering, Bachelor's degree, Electrical engineering at Sharif University of Technology
Persian, English